The future of in-house legal is built on technology designed to shape what happens next. That distinction sounds like marketing language. It is not. It describes a genuine architectural difference in how legal platforms are built, what they do with the data that flows through them, and whether the value of the investment grows or stays flat over time. The shift from system of record to system of work is the most consequential technology decision a GC makes when building the legal function of the future. Understanding the difference, recognising which model your current platform represents, and knowing what the transition requires are the starting points for any serious legal technology strategy.
The system of record model
What most legal platforms are actually built to do
What it is
A system of record is fundamentally a structured database with a legal interface. It captures what happened: matters are logged, contracts are stored, approvals are documented, and the archive of everything the legal function has touched is maintained in a searchable, auditable form.
This represents genuine and significant progress over the default non-system that most enterprise legal teams are actually running on, which is a combination of shared inboxes, SharePoint folders, WhatsApp messages to the GC, and external law firm invoices. Replacing that with a structured system of record creates real value: visibility into what is in the queue, reporting on legal output, and the ability to find things that would previously have been lost in an email thread from three years ago.
But a system of record has a structural ceiling, and that ceiling is reached faster than most GCs expect.
Why it stalls teams
The structural limitation of the system of record model is one that no amount of feature development can overcome: the 500th contract review is no smarter than the first.
The platform stores what happened. It does not learn from it. Every new matter begins from a generalised baseline, even if the same counterparty, the same clause pattern, and the same risk profile has appeared dozens of times before. The institutional knowledge that experienced lawyers have accumulated over years of reviewing contracts, every non-standard position they have spotted, every escalation pattern they have established, every risk they have learned to flag, sits in people's heads rather than in the system.
When those people are on leave, unavailable, or have left the organisation, the knowledge goes with them. The system of record captures the outcome of their decisions. It does not capture the reasoning that produced them.
What high performing teams do
Teams that have recognised the system of record ceiling have begun asking a different question of their legal technology: not just can it store what happened, but can it learn from what happened and apply that learning to what happens next?
-
Audit your current platform against the compounding question. Ask your vendor: after twelve months of use, how does the system apply what it has learned from past matters to new ones? If the answer is primarily about search and retrieval, you are operating a system of record.
-
Map the knowledge that currently lives in people's heads. The most valuable institutional knowledge is typically undocumented: standard positions on common clause types, escalation criteria for different risk levels, the commercial context that makes a particular deviation acceptable or unacceptable. Document it, and then evaluate whether your current platform can systematically apply it.
-
Assess the cost of knowledge loss. When a senior lawyer leaves, how long does it take for a new team member to develop equivalent judgment on the matters they handled? That gap is the cost of institutional knowledge living in people rather than systems, and it is quantifiable.
The system of work models
What it means for AI to learn your organisation
What it is
A system of work is different in its foundational design. It does not just capture what happened. It uses what happened to inform what happens next, systematically and at scale.
Every matter that passes through a system of work contributes to the institutional context. Clause positions accepted or rejected in past contracts inform the AI's assessment of the next one. Escalation patterns from previous matters train the triage logic for future ones. Advice that proved accurate and was acted on successfully becomes part of the knowledge base that responds to the next similar question. The system gets better at your organisation's specific legal work with every matter it touches.
After twelve months of consistent use, a system of work is materially better at your work than it was when it started. That is the compounding advantage, and it is what separates legal functions that are building permanent capability from those that are buying temporary speed.
Why this distinction matters for ROI
The ROI profile of a system of record and a system of work are fundamentally different. A system of record delivers near-term efficiency gains that plateau: the gain from having organised, searchable matter data is real, but it does not grow over time. The value of your investment in year three is roughly equivalent to the value in year one.
A system of work delivers a different profile: gains that grow. The value in year three is substantially higher than the value in year one, because the system knows more, applies more, and requires less human intervention for the work it has seen before. Teams that switch from a system of record to a system of work in year two or three of their technology investment typically describe the compounding effect as the most striking and unexpected benefit of the transition.
What high performing teams do
-
Evaluate platforms on context accumulation as a primary criterion, not a secondary one. Most legal technology evaluations focus on features, UX, and price. The question of how the system learns and applies organisational context is often asked late or not at all. Make it the first question.
-
Run a twelve-month value trajectory exercise before procurement. Estimate the value the platform will deliver at month one, month six, and month twelve. If the numbers are roughly equal across all three time points, you are looking at a system of record. If they grow, you may be looking at a system of work.
The four layers of organisational context
The architecture that makes a system of work possible
What it is
The context that a genuine system of work accumulates operates across four distinct layers. Understanding these layers helps evaluate whether a platform is genuinely building institutional knowledge or performing a sophisticated version of search and retrieval.
The system layer captures organisation-wide settings: risk posture, approval frameworks, entity structure, regulatory environment, and the guardrails that apply across all matters regardless of type. This is the foundation on which everything else sits.
The function layer captures how the specific legal team operates: the scope of work the team handles, how matters are triaged and routed, escalation paths, team playbooks for different matter types, and the allocation of work between lawyers and AI.
The service layer stores the substantive knowledge: clause libraries, standard contract templates, the positions the team takes on common issues, the precedents established across dealings with specific counterparties, and the regulatory interpretations the team has established for its specific industry context.
The personalisation layer captures individual and team preferences: how each lawyer likes to receive information, the format appropriate for different audiences, the level of detail needed at different stages of a matter.
Why it stalls teams
Most platforms operate at one layer, typically the system layer, and describe that as context. A platform that has captured your data residency requirements and approval workflow is not the same as a platform that has learned your standard positions on indemnification clauses and applies them automatically to new contracts. The distinction matters enormously for the practical value the system delivers.
What high performing teams do
-
Request a specific demonstration of each layer during vendor evaluation. Ask vendors to show, with a real-world scenario, how the system applies function-layer context, specifically the team's escalation patterns, to a new matter. Then ask how service-layer context, specifically a clause position established in past contracts, is surfaced in the review of a new one.
-
Map your organisation's knowledge to the four layers before procurement. Identify what is currently documented at each layer, what is undocumented but exists in people's heads, and what is genuinely absent. The gaps in your current knowledge map are what the system of work needs to fill.
-
Build the layers sequentially. Teams that try to populate all four layers simultaneously tend to produce shallow coverage across all of them. Teams that build the system layer thoroughly, then the function layer, then the service layer, produce a system that genuinely knows their organisation by month twelve.
Source: Plexus Future-Ready General Counsel 2026 Survey. External citation: Thomson Reuters Future of Professionals 2025.
Ready to see what a system of work looks like? Explore the Plexus platform or book a demo.
.svg){kind=small}
-1.png)
Meet Plexus Counsel, your new AI-powered digital lawyer > See it in action
.svg){kind=small}
